[Enhancement]: Support a way to customize agents' behavior more deeply #6498
Replies: 1 comment 3 replies
-
I think some help defining these things more specifically will help me understand your needs. Furthermore, you should review the current scope of agent capabilities because they are highly configurable: https://www.librechat.ai/docs/features/agents#agent-chain This newer feature of agent chaining helps create a "pipeline" so to speak, largely relying on your customization of instructions, tools, and resources attached.
You shouldn't be doing this any longer, you can use actions, as you said, as well as MCP servers: All that said, improving agents is a core priority of mine, and I will showcase in an upcoming video just how capable they already are. |
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
-
What features would you like to see added?
Some ways to customize agents behavior more easily
More details
I think the current definition of agents lacks flexibility. I know there are some solutions, such as building custom plugins or actions, but they may not be sufficient. Each time I need to create a new plugin, I have to modify some project files and map them into Docker container to overwrite the original files.
Maybe we should provide a more flexible approach, such as implementing pipeline like Open Web UI or integrating Dify as a backend? Dify offers a complete solution for customizing agents and workflows. I think integrating it natively into LibreChat would be a huge advantage.
Which components are impacted by your request?
No response
Pictures
No response
Code of Conduct
Beta Was this translation helpful? Give feedback.
All reactions